Akkariin/youtube-dl
1
0
Fork 0
mirror of https://github.com/l1ving/youtube-dl synced 2025-03-13 07:17:29 +08:00
Code Issues Releases Wiki Activity

[Bandcamp] Extract additional fields

Browse Source 
* [bandcamp] Add additional extraction fields
This commit is contained in:
gkoelln 2017-01-25 10:49:24 -06:00
parent 1f393a3241
commit cc2e170d86
1 changed files with 31 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
youtube_dl/extractor/bandcamp.py
View File

@ -16,6 +16,7 @@ from ..utils import (
int_or_none, int_or_none,
parse_filesize, parse_filesize,
unescapeHTML, unescapeHTML,
unified_strdate,
update_url_query, update_url_query,
) )
@ -47,6 +48,18 @@ class BandcampIE(InfoExtractor):
mobj = re.match(self._VALID_URL, url) mobj = re.match(self._VALID_URL, url)
title = mobj.group('title') title = mobj.group('title')
webpage = self._download_webpage(url, title) webpage = self._download_webpage(url, title)
release_date = self._search_regex(
r'"?(?:album_)?release_date"?\s*?:\s*?"(.*?)"',
webpage, 'release date', fatal=False)
release_date = unified_strdate(release_date)
album = self._search_regex(
r'album_title\s*:\s*"((?:\\.|[^"\\])+?)"',
webpage, 'album', default=None)
artist = self._search_regex(
r'artist\s*:\s*"((?:\\.|[^"\\])+?)"',
webpage, 'artist', fatal=False)
if release_date:
release_year = release_date[:4]
m_download = re.search(r'freeDownloadPage: "(.*?)"', webpage) m_download = re.search(r'freeDownloadPage: "(.*?)"', webpage)
if not m_download: if not m_download:
m_trackinfo = re.search(r'trackinfo: (.+),\s*?\n', webpage) m_trackinfo = re.search(r'trackinfo: (.+),\s*?\n', webpage)
@ -76,7 +89,14 @@ class BandcampIE(InfoExtractor):
'id': track_id, 'id': track_id,
'title': data['title'], 'title': data['title'],
'formats': formats, 'formats': formats,
'release_date': release_date,
'duration': float_or_none(data.get('duration')), 'duration': float_or_none(data.get('duration')),
'track': data.get('title'),
'track_number': int_or_none(data.get('track_num')),
'track_id': track_id,
'album': album,
'artist': artist,
'release_year': release_year,
} }
else: else:
raise ExtractorError('No free songs found') raise ExtractorError('No free songs found')
@ -86,6 +106,10 @@ class BandcampIE(InfoExtractor):
r'(?ms)var TralbumData = .*?[{,]\s*id: (?P<id>\d+),?$', r'(?ms)var TralbumData = .*?[{,]\s*id: (?P<id>\d+),?$',
webpage, 'video id') webpage, 'video id')
track_number = self._search_regex(
r'"track_num":(?P<track_number>\d+),',
webpage, 'track number', fatal=False)
download_webpage = self._download_webpage( download_webpage = self._download_webpage(
download_link, video_id, 'Downloading free downloads page') download_link, video_id, 'Downloading free downloads page')
@ -145,9 +169,15 @@ class BandcampIE(InfoExtractor):
'title': title, 'title': title,
'thumbnail': info.get('thumb_url'), 'thumbnail': info.get('thumb_url'),
'uploader': info.get('artist'), 'uploader': info.get('artist'),
'release_date': release_date,
'artist': artist, 'artist': artist,
'track': track, 'track': track,
'formats': formats, 'formats': formats,
'track_number': int_or_none(track_number),
'track_id': video_id,
'album': album,
'artist': artist,
'release_year': release_year,
} }
@ -233,5 +263,6 @@ class BandcampAlbumIE(InfoExtractor):
'uploader_id': uploader_id, 'uploader_id': uploader_id,
'id': playlist_id, 'id': playlist_id,
'title': title, 'title': title,
'album': title,
'entries': entries, 'entries': entries,
} }